Mathematical Reasoning on MATH500 (Accuracy and Output Length)
88.7Accuracy0.7 × RoPE theta × 16 + 0.3 × 1M-Model
Evaluation results
| Method | Links | ||
|---|---|---|---|
| 0.7 × RoPE theta × 16 + 0.3 × 1M-ModelOperation=0.7 × RoPE theta × 16 + 0.3 × 1M-Model, recipe=LCTMerge2025.05 | 88.7 | 5,731 | |
| RoPE theta × 16Operation=RoPE theta × 162025.05 | 88.2 | 5,501 | |
| RoPE theta × 1Operation=RoPE theta × 12025.05 | 85.04 | 1,584 | |
| RoPE theta × 1Operation=RoPE theta × 12025.05 | 81.88 | 2,037 | |
| 0.7 × RoPE theta × 16 + 0.3 × 1M-ModelOperation=0.7 × RoPE theta × 16 + 0.3 × 1M-Model, recipe=LCTMerge2025.05 | 74.12 | 2,228 | |
| RoPE theta × 16Operation=RoPE theta × 162025.05 | 65.16 | 4,579 |